(ESMFold2) Folds molecular complexes containing proteins, dna, rna, and ligands. Defaults to esmfold2-fast-2026-05 if no model is given

You may have access to additional, private models as well. These will be visible on the Rate Limits page under the Developer Console

Disclose potential sequences of concern. For approved users, such sequences will not go through additional safety filtering. Reach out if you are interested in using this.

Input sequence to be folded; '|' separates chains. The server names the chains A, B, C in the order they appear.

Number of trunk loops for iterative refinement.

Diffusion ODE solver steps. Lower for speed, higher for quality.

Dropout probability on LM pair embeddings. When > 0, dropout is applied.

Fraction of sequence residues randomly masked before the PLM backbone. If not provided, defaults to 0.1 for ESMFOLD2_FAST and 0.0 for ESMFOLD2

Number of MSA rows randomly subsampled each loop. Set to null to disable (sets msa_subsample_at_inference to False).

Fraction of MSA columns randomly masked in non-query rows for inference-time diversity.

Whether to include distogram predictions in the response.

Whether to include Predicted Aligned Error (PAE) matrix in the response.

Whether to include pair-chain IPTM predictions in the response.

Whether to include sequence and pair embeddings in the response.

You may have access to additional, private models as well. These will be visible on the Rate Limits page under the Developer Console

ISO formatted date time

Indicate that the returned sequence may be of potential concerns.

Folded protein sequence in coordinate order. Always populated by /fold; null for all-atom folding.

Predicted TM score of a protein or molecular complex. Usually produced by structure decoder.

Interface TM score of a protein or molecular complex.